论文链接: https://arxiv.org/pdf/1506.02640.pdf 代码下载: https://github.com/gliese581gg/YOLO_tensorflow Abstract We present YOLO, a new approach to object detection.Prior work on object detection repurposes classifiers to perform detection. Instead, we frame
Yolo:实时目标检测实战(上) YOLO:Real-Time Object Detection 你只看一次(YOLO)是一个最先进的实时物体检测系统.在帕斯卡泰坦X上,它以每秒30帧的速度处理图像,在COCO test-dev上有57.9%的mAP. 与其他探测器的比较,YOLOv3非常快速和准确.在0.5 IOU处测得的mAP中,YOLOv3与焦距损失相当,但速度快了约4倍.此外,可以轻松地权衡速度和准确性之间的简单改变模型的大小,无需再训练! COCO数据集的性能 How it works
转自:https://blog.csdn.net/KKKSQJ/article/details/83587138 original Based on keras-yolov3, understanding of the principle and code details October 31, 2018 17:37:43 Aries seven seven seven reading number: 2917 This article GitHub source code : https:/